% Things to do in Tex
	% 1. Add \begin{table} and \end{table}
	% 2. Replace \hlines with toprule middlerule & bottomrule
	% 3. Bold the top row
	% 4. Add a caption and label
	% 5. Center
	% 6. Add a note
	% 7. Add scalebox
	% 8. Remove the vertical lines between ll and rr
	% 9. Flip Endline-Only and Baseline-Only Rows
	% 10. Add notes in threeparttable format

\begin{table}[h]
\caption{Attrition Table}
\label{tab:Attrition}
\begin{center}
\begin{threeparttable}
\renewcommand{\TPTminimum}{\linewidth}
\scalebox{.9}{
\begin{tabular}{ll|rr}
\toprule
\textbf{Variable} & \textbf{Attrition Category} & \textbf{Estimate} & \textbf{p-value} \\
\midrule
Male & Endline Only & -0.029 & 0.30\\
Male & Both Surveys & 0.002 & 0.97\\
Age & Endline Only & -3.504 & 0.08\\
Age & Both Surveys & -0.317 & 0.70\\
College Educated & Endline Only & 0.014 & 0.57\\
College Educated & Both Surveys & -0.025 & 0.47\\
Catholic & Both Surveys & -0.008 & 0.59\\
Years Served & Endline Only & -3.030 & 0.10\\
Years Served & Both Surveys & -0.310 & 0.73\\
Lives in Sorsogon & Endline Only & 0.009 & 0.87\\
Lives in Sorsogon & Both Surveys & 0.097 & 0.01\\
NPA Presence 2015 & Endline Only & 0.016 & 0.87\\
NPA Presence 2015 & Both Surveys & 0.042 & 0.23\\
Crime Rate & Endline Only & 0.023 & 0.72\\
Crime Rate & Both Surveys & -0.029 & 0.24\\
\midrule
Attitude Idx (endline) & Endline Only & -0.055 & 0.20\\
Trust Idx (endline) & Endline Only & -0.124 & 0.26\\
Empathy Idx (endline) & Endline Only & -0.052 & 0.31\\
Account Idx (endline) & Endline Only & 0.148 & 0.00\\
Corrupt Idx (endline) & Endline Only & -0.226 & 0.00\\
Attitude Idx (baseline) & Baseline Only & 0.067 & 0.24\\
Trust Idx (baseline) & Baseline Only & -0.101 & 0.28\\
Empathy Idx (baseline) & Baseline Only & -0.001 & 0.99\\
Account Idx (baseline) & Baseline Only & 0.057 & 0.31\\
Corrupt Idx (baseline) & Baseline Only & 0.074 & 0.31\\
\bottomrule
\end{tabular}}
\end{center}
\begin{tablenotes}[flushleft]
\small{\item Each row reports the estimates and p-values from regressions that predict covariates based on officers' attrition status. These estimates are calculated without respect to the blocks unlike in Table \ref{tab:BalanceTable}, since only treatment was assigned with respect to the blocks. Attrition Category is a categorical variable with the following levels: \emph{Baseline Only} (officers who were assigned in Sorsogon at baseline but attritioned before endline), \emph{Endline Only} (officers who were newly assigned in Sorsogon after baseline), and \emph{Both Surveys} (officers who were assigned in Sorsogon through the duration of the experiment). The comparison group for the demographic variables is \emph{Baseline Only} (with the exception of Catholic, which was measured only at endline). The comparison group for all survey indices is \emph{Both Surveys}. Because the endline survey indices (last four rows) are measured post-treatment, these regressions are subset to officers in the control group.
\end{tablenotes}
\end{threeparttable}
\end{table}